This MCP server exposes Miro API endpoints as tools that AI agents can invoke through the MCP protocol. It automatically generates tool definitions from the Miro OpenAPI specification, enabling programmatic manipulation of a Miro board, including creating, updating, and querying various board items such as sticky notes, shapes, text, images, cards, and app cards. The server runs a FastMCP-based interface and exposes endpoints corresponding to board, item, and media operations, so agents can compose sequences of actions to manage boards, items, connectors, and documents.

To use it, configure your environment with a Miro API token and the target board ID. Start the MCP server (via the configured command, for example using npx luotocompany-mcp-server-miro). Once running, connect your MCP client or agent to the server’s endpoint (e.g., the provided SSE URL) and begin discovering and invoking the available tools. The tool set is categorized by entity type (board, items, connectors, documents, embeds, images, app cards), making it straightforward to navigate operations such as creating a new sticky note, updating a shape, creating a connector between items, or retrieving a list of items with filtering options.

Prerequisites:

• Node.js installed on your machine
• Access to a Miro account with API capabilities
• A Miro API token and a target Miro board ID

Installation steps:

1. Clone the repository (or prepare your environment to fetch the MCP server package): git clone https://github.com/yourusername/mcp-server-miro.git cd mcp-server-miro

2. Install dependencies: npm install

3. Create a .env file in the root with the required environment variables (or supply them via your hosting environment): MIRO_API_TOKEN=your_miro_api_token MIRO_BOARD_ID=your_miro_board_id

   Optional, port for the server if you want to override the default

   PORT=3000

4. Start the server locally (example scripts): npm run dev or npm run build npm start

Note: If you prefer running via npx as shown in the MCP config, ensure you have network access to fetch the package luotocompany-mcp-server-miro and provide the required environment variables when starting the process.

Tips and notes:

• Ensure your Miro API token has permissions for the board you intend to manage.
• The server exposes a wide range of tools mapped to Miro operations. Use the provided tool names (e.g., get_items, create_sticky_note_item, update_shape_item) to construct your agent prompts.
• If using Docker, follow the container-specific setup in the repository, ensuring environment variables are passed into the container.
• The PORT variable is optional; if not specified, the server may default to 8899 or another port defined by the implementation.
• For large boards, consider using pagination and filtering options when listing items to manage performance and API limits.
• Check CORS and SSE endpoint configuration if your agent connects from a different origin or environment.